18 December 2025
Some quotes are worth sharing. Roman Mars is the host of the podcast 99% Invisible. In the episode celebrating their 15th year, Roman was asked some questions from staff and listeners. One question was:
Q: What is a design-related hill you are willing to die on?
The one that's sticking with me, and this is going to be kind of abstract and I don't know if this is going to come off well, is the march to make things more and more efficient makes the world a worse place.
I think of this in terms of advertising. The idea that we were going to efficiently measure how effective advertising was through clicks and eyeballs and stuff erased all of the extra money that made all the journalism and all the pop culture that you cared about in the 20th century. It made it all possible. The inefficiency of the advertising system made everything good in this world. I think that the idea that you are trying to get things as efficient as possible is actually a terrible, world-destroying idea.
Like, the most efficient restaurant is a ghost kitchen that has no storefront. That's inefficient because it could be empty sometimes. And it's a ghost kitchen that just ships you a thing, and has an underpaid delivery person that brings it to your door, and you never leave.
This is stripping away all the goodness of the world and of cities. I think efficiency is absolute garbage and that is the design-related hill I am willing to die on.
I feel you should always be allowing for great deals of inefficiency to make a nice designed city, a nice design system, and make it work. I really hate the focus on efficiency.
Not only does it destroy all these good things, it takes money and gives it to the worst people—the platform creators and the tech people—instead of all the loose, empty, change and tips and things. It's not like things are cheaper or things are better. Money is being transfered to the wrong people instead of creators and people who make the world a better place through community and creation.
"Yea, a more efficient world is a place where more people are left out, essentially. The more inefficient, the more people are included." ~ 99PI producer Vivian Le.
Totally. You need friction, you need creation. That open freedom of an inefficient system where money sloshes around inside of it. These frictions and inefficiencies are what make everything good about the world. I want to hear nothing about how good efficiency is or the Department of Government Efficiency. Nonsense. That is the worst world we are building when we do efficent things.
My first thought went to the insertion of analytics into the world of sports. This story often begins with the book Moneyball: The Art of Winning an Unfair Game by Michael Lewis, later adapted into a movie. An even earlier beginning is the book Baseball Perspectus by Earnshaw Cook. The biggest takeaways are simple enough, that batting average wasn't as important as on base percentage (OBP). Getting on base any way possible was more important than the classic batted ball into the shallow outfield for a single. But the findings evolved until everyone concluded the most efficient way to score runs was solely through the home run. In addition, walks are important, but strikeouts are not. Add the proliferation of defensive shifts, and we arrived at the "three true outcomes" of baseball: the strikeout, the walk, and the home run. The first two are undramatic. The last provides about four seconds of fun within a three-hour game.
Basketball began with just a peach basket on each end. The tallest players closest to the basket scored the most points. The game was just throwing the ball down the court to the player near the basket. They would quickly score or a scrum would happen. To make the game more exciting, the ABA invented the three-point line, incentivizing a wider range of scoring options. When analytics arrived, it turned out only dunks and three-point shots were even worth attempting. Since then, the only innovation has been how strictly teams adhere to this mantra. As veteran players retire, so too does the teamwork promoted by classic point guard play and the long-range two-point shot.
American football has had its own influx of statistics like DVOA (defense-adjusted value over average) and expected points gained from each play. There is also the infamous two-point conversion and fourth down charts, informing coaches of when to try each given the situation. Running backs are less important than ever as offenses have been converted to a "throw, baby, throw" mantra.
Soccer has been introduced to expected goals (xG), expected assists (xA), progressive passes, and the like.
But there has also been pushback. Analytics alone ignores human factors. On the field, how do you account for a psychological hurdle? Is a certain nemesis in a player's head? Did a player stay out too late or get sick overnight? Are half the players newly acquired and haven't been able to syncronize their actions yet? Did a coach say something offensive in practice recently and there's a simmering feud? Do your players simply have a unique set of skills? As a coach, do you trust them?
Recently, a baseball pitcher was trying to wrap up a no-hitter, but gave up a home run instead of getting the last out. The pitcher was pulled since they were no longer going for the resume item anymore. No worries, the team was still ahead 3-1. But a double, hit by pitch, wild pitch, walk, walk, and single later, the team lost 4-3. The starter couldn't have gotten one more out? I know about 2004 Pedro Martinez, but that's not the only possible outcome. If you pull starters earlier, you are requiring all your relievers to not have an off-night.
They say throwing the football can either result in a catch, an incomplete pass, or a pass interference call by a referee, and two of those three are good outcomes. But any football player will tell you there is nothing more demoralizing than a team running on you and your team not being able to stop them.
When only 5 players are on the court at once, there is statistically a big benefit to your best player having the ball all the time. This "hero ball" style characterized James Harden on the Houston Rockets, Luca Doncic on the Dallas Maverics, and a few others. A statistical argument was made that 2015-2016 Steph Curry maybe should just take all the shots. But if one player is doing all the work, the other four players are likely to drop their effort. It ruins the vibe. Ruins the fun. Players stop trying as hard on both offense and defense. Good defense is primarily about a continual emphasis on effort. If effort is thankless, effort wanes.
Leagues are dealing with frustrated fans. If all the teams play the same way, that's boring. If all the excitement happens in an event with no build up and is over quickly, that's boring. We don't just want to see who is the best at a single task. We want to see a clash of styles. We want to see athletes be more than just precise. We long to see the unexpected. We want drama to build and peak.
There is also a methodological issue. Many summary statistics and analyses are based on averages of all players or teams from all years. Stats like WAR, PER, xG, and the fourth-down chart are all examples. But neither the 2016 Golden State Warriors (69 wins) or 2023 Detroit Pistons (14 wins) have the same likelihood of success as the average outcome the stat found. This is true of every sport.
Averages are guides, not truths.
Humans are the distribution around the average. That's where the interesting is.
The average starting pitcher may indeed do much worse the third time through the rotation, but that doesn't account for your starting pitcher today, what your bullpen's gone through, who this opponent is, and planning for the next few games. Success in a 162 game season is not achieved the same way as in a 7-game playoff series.
In basketball, your best player may be the winning option in the regular season. "Hero ball" does work, on average, over the course of the regular season. But it's easy to plan for when repeatedly playing a team in a playoff series. "Hero ball" has a bad playoff record. Suddenly, if your best player gets defended well or decides to "Harden Quit" during the playoffs, the rest of the players haven't had to carry the team during the season. They won't be able to make up for that lack of experience during the heat of the playoffs. Success in an 82 game season is not achieved the same way as in a 7-game playoff series.
The fourth-down chart may say to go for it on fourth-and-four, but if your team is exhausted and demotivated, then what are we doing? Playoff rounds that are one game can be easily messed up in one play.
Supposedly, there are people who enjoy talking to others in their communities and identifying structural problems that need addressing. People with potential solutions for those problems run in an election and the community chooses who likely has the best ideas. The elected officials from all the disparate communities then assemble to see if there are common solutions that address systematic problems across the nation. Obviously, communities are different, so some negotiation and smoothing the rough policy proposal edges will be required.
Of course, that lovely picture presumes that news isn't constantly travelling between communities. There have been national and international news for as long as there have been nomadic tribes. So that ideal scenario has never reflected reality. Getting stuff done and turning policies into law has always required coalition building and deferring to the leaders among the elected officials in hopes of your personal policy of interest getting promoted eventually.
The computer and analytics age then transformed politics. Democracy as a form of government was created so the policies that should best serve our communities would be prioritized. Policies and communities should remain the force that moves elected leaders. But politicians have used manipulative tactics to gain advantages as long as there have been leaders. Modern computing and analytics provided elected leaders, political parties, and their funders with insights into election results and the voters themselves. Those insights then motivate leaders to mess with the rules governing elections in increasingly precise ways to best ensure their party wins elections. Parties are fiddling with laws about who can vote, and what is needed to be able to vote, to tilt the results of who will vote toward the demographics who have voted for their party in the past. Perhaps the most notorious type of fiddling with the rules is the process of gerrymandering. For the politicians, as long as you have the majority, you can do what you want. So you do whatever you can to take the uncertainty out of the next election.
The result is that policies and communities are no longer the force that moves elected leaders. Elections cannot serve as the voice of the community when their rules are shaped to keep the current party in power. When it seems politicians are just trying to ensure electoral victory, it separates the community from the elected leadership. If successful, it's other leaders and party infrastructure that become the force that moves elected leaders. The policies the leaders then want to prioritize are not what the community needs.
It is certainly the reality that communities do still elect leaders, controlling what party is in power. But even felt separation from the elected leaders and parties, perceived disconnect between policies that leaders persue and what we need, appears to me to leave us feeling disenchanted with the leaders we voted for and desparate to remove leaders we did not vote for. If elections are even perceived to not serve as the connection between communities and leaders, then governemnt does not feature community representatives in the way originally hoped for.
Monopoly law is antitrust law. Many charities today are run as "trusts," where "trustees" are leaders who manage the operations for the benefit of the charity. Pensions are often run as "trusts," where the employer delegates the management of the money to trustees who act on behalf of and for the benefit of the employees.
If someone wanted to do business in the 1800s, they had to set up a separate company in each state. John Rockefeller set up a trust so Standard Oil would have a single management structure to control all the companies in the 38 states that existed in 1882. Nine trustees had majority control over 40 corporations. This centralized control also meant they could accumulate and then abuse their market power as a monopoly. The laws passed to guard against "anti-competitive business practices," or to assure multiple companies could always reasonably compete in all markets, are collectively known as antitrust law.
The US Congress passed a law in 1890 against monopolization and using corporate mergers to gain a monopoly. The Supreme Court used it to break up Standard Oil into 34 separate companies, arguing it controlled too much of the market and unreasonably restrained trade.
Economists in the 1970s began to argue for changes to these laws with theories regarding what corporate actions were truly anti-competitive. These economists began arguing that as long as monopolies benefitted the consumer through lower prices, then a monopoly should not be broken up. Just because there is a monopoly doesn't mean there could be the potential for competition, they reasoned. Only coercive monopolies that can raise prices without the threat of competition are bad.
With antitrust law now focused on a single measure, companies can now use analytics to determine how efficient they can be to maximize profits—maximize price gouging and coercive monopolization—while also being able to argue they are keeping prices low for the consumer. Holistic examination of corporate policies are ignored in the name of efficiency. Therefore, we may not have single company monopolies, but we still have fewer choice than ever. Aviation is dominated by four companies. Car manufacturing is dominated by three companies. Car rental is dominated by three companies. Grocery stores are dominated by six companies. Mobile phones are dominated by three companies. Computers are dominated by two companies. We have "all but monopolies."
It appears universities across the nation are in a funding crisis, compounded by a lack of incoming students providing tuition.
Low enrollment. Low producing. Low graduation rates. Less in-demand careers. Data-informed. Metrics-based. The numbers are the numbers.
But why is enrollment the metric? Tuition and fees are often the single largest source of income for universitiesAccording to the latest operating budgets: 40% at West Virginia University. 27% at the University of Oklahoma. 22% across the Universities of Wisconsin, ranging from 19-35% across the 13 universities individually. 20% across all University of Nebraska locations. 11% across Florida public universities., but it's still only a quarter of income at the institutions listed, on average.
For their part, the states are generally continuing to fund public universities. The National Education Association reported in 2025 that state appropriations for higher education are higher than pre-pandemic levels for 38 states, after adjusting for inflation. The report says this indicates a "more robust and sustained reinvestment in public higher education." Looking further at the State Higher Education Finance data used for the report, total state appropriations across all 50 states has increased fairly steadilyExcept for 2021 after COVID. since 2013, following five years of cutbacks following the 2009 recession. Behind that national trend, increases in state appropriations only came from 35 statesAlaska (-31%), Arkansas (-23%), West Virginia (-23%), Oklahoma (-20%), Wyoming (-20%), Iowa (-16%), Kentucky (-8%), Mississippi (-8%), Indiana (-6%), Louisiana (-5%), Ohio (-5%), Illinois (-2%), New York (-2%), North Dakota (-2%), Pennsylvania (-1%). during that time, again relative to inflation.
The purpose of education is to learn skills and knowledge that will help students contribute to and participate in society. Academic programs beyond the K-12 education, especially public colleges and universities, are meant to provide specialized skills and knowledge specific to a career path that will contribute to the needs and prosperity of the state.
What is frustrating is the cuts aren't being made in a way that reflects the future needs or prosperity of each state. Budget deficits are being met with a single measure—how much tuition is being brought in by each program.Accusations of political motivations are outside the scope of this post. Both humanities and STEM programs have been cut. There isn't an examination of trends in what the state will need compared to current enrollment. There isn't an analysis of what the university system as a whole can support. The measure is effectively whether each degree program is self-sustaining, at least regarding their part of expected tuition.
Nobody is saying the state won't need future engineers, geographers, historians, or local entertainment. If the programs aren't self-sustaining, they are being cut. There seems to be little effort to try to use profitable programs to help sustain low-enrollment programs.
Research is a process. Teaching is a process. Learning is a process. All of them are long-term investments. None are single-year profit margin enterprises. What looks like economic efficiency is not a replacement for planning.
How does a pastor know they are doing their job?I'm sure there is truth to this for other faiths, but this one is mine and so is what I'm most passionate about.
In any other line of work, you are designing wigits, making wigits, moving wigits, selling wigits, fixing wigits, cleaning wigits, entertaining people, doing a task for people, or guiding people. But a member of clergy is doing these jobs, or delegating a team to do these jobs, so that people in your care have a greater relationship with their neighbors and with God.
Relationship strength is kinda hard to quantify. Surveying your audience kinda dampens the vibe being built, which defeats the purpose. There are also methadological pitfalls to worry about. I'm pretty sure Matthew 4:7 means God isn't a fan of filling out surveys.
The other big dynamic is that Christianity is a team sport to be done together. To ensure the best team, clergy should delegate tasks to others while casting a vision that the community will be successful by following the plan. Well, the community would like some reassurance that what they are doing is indeed worthwhile. So the issue is bigger than whether the clergy are doing their job. Is the group, following the clergy's vision, doing their job?
If only there was something to point to! After all, everyone in the community has "normal" jobs with annual performance reviews. All their tasks can be quantified: How many wigits did you design, make, move, sell, fix, or clean? How many people did you entertain, guide, or do tasks for? How much money was made? Surely, if the clergy could quantify what the group is doing, reflecting what the community is experiencing, the big number would just increase the prideI'm not worried about the sin of pride, here. I'm not trying to use Christianese here. Let's not overthink things. in their work! We're making a difference! People are used to thinking in numbers. That pride would be in their work, the community they are with, and in the clergy's vision.
So the measurements begin. How many people are coming to church? Is attendance increasing? How much is the community donating? How many members attended the event? How many people were served? How many people heard the gospel? For how long were we talking to people? How many salvations were there last month? How many baptisms? How many spiritual gifts did people experience or demonstrate? How many people are speaking in tongues for the first time? How many miraculous healings were there? How many exorcisms?
You can just see how denominations get characatured. Sure, denominations are primarily created for theological reasons. I'll let theologians argue about infant baptism, Calvanism, or cessationism. But you can see how denominations fall into the traps of making everything about one number—the size of their congregation (megachurches), the number of salvations (Baptists), regular miracles (charasmatics), political power (Christian nationalists), donations (prosperity gospel), or website quality (non-denominational).
To measure is to miss the goal.
By definition, efficiency is getting the maximum productivity with the minimum wasted effort or expense. The flaw of numbers, analytics, and measurements is they do not factor in people. Measurements are great at telling us about the past and about what has been "typical." They are only indicators of the future so long as the people and our systems do not change. But people and systems do change.
In a business, efficiency means people do not gather or commune. Individuals minimally interact with an employee and return to whatever they were doing, likely by themselves. The employee is in a repeatable job that takes no training, yet is one of the very few people between the owner, CEO, or board and the employees. Efficiency allows the most amount of money to be transferred from the customer to the owner, CEO, or board. Efficiency is great for the owner, but it’s bad for people.
When there is only one efficient way to win, sports become boring. Variety and built drama are taken out of the games. Efficiency is mostly great for winning when it doesn’t matter, but it’s bad for people.
When there is an efficient way for a party to be elected, then leaders need to do the least amount of campaigning and communties become disconnected from the leaders they elect. At worse, this prevents passage of the policies that communities need to best thrive. Communities don’t like the leaders they elect and despise the leaders they didn’t elect. Efficiency is mostly great for the politician, but it’s bad for people.
When there is only one measure of a coercive monopolly, companies only have to argue their number is on the good side of the threshold. The supposed efficiency of a low shelf price ensures nobody will focus on their holistic actions. This hurts the holistic economy, and prevents the continual creation and innovation of small business. Efficiency is mostly great for the company owners, but it’s bad for people.
When there is only one measure of whether a major should be offered, public universities cease to provide for the holistic future needs of the state. Instead, we end up with a glut of professionals in programs that were historically deemed economically self-sustainable in decades past. Efficiency looks mostly great to budget managers, but it’s bad for people.
When church clergy rely on metrics to prove their success, and their congregation’s success, they start trying to maximize their favorite number instead of providing the holistic service demanded by their faith. Efficiency looks mostly great to the clergy, but it’s bad for people.
Large language models have the same issue. The algorithms tell us the average of what people know. But humanity is in the distribution around the average and expertise is in the specifics that support the bottom line, generic statement. Large language models create text efficiently. There is no need for the inefficiencies of navigating relationships and trust. As a result, large language models do not provide truth, authenticity, or perspective that come with the inefficiency of a lifetime of experiences, failure, and learning.
Sports have already figured out that adherance to analytics alone during the regular season goes wrong in the playoffs precisely because they ignore the human factors involved. Athletes learn how to defend what analytics said was the best way to score. Leagues have changed their rules to ensure their sport remains entertaining.
Voters become frustrated when it seems policies are taking away their agency. We need a mechanism that will maximize a politician’s connection with their community, despite the impacts on their job security.
The good of the whole is harmed when business, university, and church leaders favor a single reductive number. People are harmed when leaders say they are just reacting to facts, or economics, instead of planning for the good of all.
Efficiency can only be a worthy goal when it is paired with the primary long-term mission. A tool cannot itself be the goal. Statistics are a tool that can only be properly used when the operator knows the context of the project and what they will do after the tool is no longer needed. That context always involves people.
Inefficiencies mean more people are involved, and help us interact. Human interaction is what helps us thrive as humans. To expand our sense of how big our team is, efficiency needs to be deemphasized as a goal and return to being a tool to be used appropriately.